Dissertation Defense of Ting Luo, April 12th, 1pm, M1170 SPH II

Post date: Apr 10, 2018 5:04:15 PM

The University of Michigan

Department of Epidemiology

Presents the

Dissertation Defense

of

Ting Luo, MPH

speaking on

Algorithms & Techniques for Studying In Vitro Oral Biofilms

Thursday, April 12, 2018

1:00 – 2:00pm

Room M1170 SPH II

All interested persons are cordially invited

Dental caries and periodontal disease affects billions of people worldwide with an estimated global prevalence of 35% and 11%, respectively. Oral biofilms that develop on tooth surfaces and within gingival crevices are a major risk factor. Disease prevention efforts are focused on controlling the overgrowth of biofilms by removal (e.g., tooth brushing), antimicrobial-containing mouth rinses, and dentifrices. A number of laboratory (in vitro) models of biofilms are used to understand how biofilms develop and their response to mouth rinses and dentifrices.

However, there are two major limitations to currently available in vitro biofilm model systems. First, there is no biofilm model system validated for the development of representative dental plaque biofilms. Second, there is no standard approach to analyze biofilm images. Current techniques rely on thresholding algorithms that are not designed for fluorescent images. Combined, these limitations can lead to differences in quantification of biofilm outcomes and thus raise questions regarding the relevance of the model system to the real-world.

This dissertation seeks to bridge the gap between current laboratory techniques and software algorithms and provide investigators complementary protocols to conduct in vitro oral biofilm studies. First, we provided a review of existing model systems relevant to modern in vitro oral biofilm research. Second, we adapted the 24-well Bioflux, the most versatile of the reviewed systems, to reproducibly grow multi-species dental biofilms seeded from human saliva. Additionally, we devised an objective imaging strategy to capture all biofilm architectural features. From the analytical perspective, we designed a novel thresholding algorithm, the biovolume elasticity method (BEM) to threshold fluorescent signal. Finally, we built a software package called Biofilm Architecture Inference Tool (BAIT) to measure core architectural features of biofilms.

In summary, this dissertation describes the modification of a 24-well Bioflux system that facilitates the reproducible development of oral biofilms. The BAIT package, featuring the novel BEM thresholding technique, enables investigators to quickly quantify an array of biofilm architectural descriptors. Understanding change and continuity in biofilm architecture after treatment can identify potential new agents of biofilm control. The work presented here represents the outcome of a combinatorial approach to redefine techniques to study oral biofilms which may be relevant to the study and analysis of all biofilms that have been fluorescently labeled.